# Scale Gymnasium

## Docs

- [Accessibility](https://docs.gym.scale.com/api-reference/desktop/accessibility.md): Get the full UI accessibility tree
- [Close Environment](https://docs.gym.scale.com/api-reference/desktop/control-plane/close.md): Terminate VM and clean up resources
- [Create Desktop](https://docs.gym.scale.com/api-reference/desktop/control-plane/create-desktop.md): Provision a new VM environment
- [Get VNC URL](https://docs.gym.scale.com/api-reference/desktop/control-plane/get-vnc-url.md): Get the VNC connection URL for a VM
- [Health Check](https://docs.gym.scale.com/api-reference/desktop/control-plane/health.md): Check if the OSWorld API server is running
- [Initialize Task](https://docs.gym.scale.com/api-reference/desktop/control-plane/initialize-task.md): Set up a task configuration in a VM environment
- [List Environments](https://docs.gym.scale.com/api-reference/desktop/control-plane/list-environments.md): View all active desktop environments
- [Reset Environment](https://docs.gym.scale.com/api-reference/desktop/control-plane/reset.md): Reset the desktop environment to initial state
- [Run Evaluator](https://docs.gym.scale.com/api-reference/desktop/control-plane/run-evaluator.md): Evaluate task completion and get a score
- [Task Status](https://docs.gym.scale.com/api-reference/desktop/control-plane/task-status.md): Check the status of a background task
- [Execute](https://docs.gym.scale.com/api-reference/desktop/execute.md): Execute a shell command or script inside the VM
- [File](https://docs.gym.scale.com/api-reference/desktop/file.md): Download a file from the VM's file system
- [Desktop Environment API](https://docs.gym.scale.com/api-reference/desktop/overview.md): API reference for Desktop VM management and interaction
- [Platform](https://docs.gym.scale.com/api-reference/desktop/platform.md): Retrieve information about the VM's operating system and platform
- [Screenshot](https://docs.gym.scale.com/api-reference/desktop/screenshot.md): Capture the current visual state of the VM desktop
- [Call Tools](https://docs.gym.scale.com/api-reference/mcp/call-tools.md): Execute a specific tool with provided arguments
- [Health Check](https://docs.gym.scale.com/api-reference/mcp/health.md): Check the health status of the MCP environment
- [List Tools](https://docs.gym.scale.com/api-reference/mcp/list-tools.md): Discover all available tools across MCP servers
- [MCP Environment API](https://docs.gym.scale.com/api-reference/mcp/overview.md): API reference for MCP server interactions
- [Show Data](https://docs.gym.scale.com/api-reference/mcp/show-data.md): Inspect the current state of an MCP server's database
- [API Reference Overview](https://docs.gym.scale.com/api-reference/overview.md): API documentation for Scale Gymnasium environments
- [Website Environment API](https://docs.gym.scale.com/api-reference/websites/overview.md): API reference for synthetic website interactions
- [Prompt Convert (Calendr only)](https://docs.gym.scale.com/api-reference/websites/prompt.md): Convert variable-based dates to absolute dates
- [Reset Session](https://docs.gym.scale.com/api-reference/websites/reset.md): Initialize or reset a session with optional data pack
- [Task Transform (Calendr only)](https://docs.gym.scale.com/api-reference/websites/task.md): Convert absolute dates to variable-based dates
- [Verifier](https://docs.gym.scale.com/api-reference/websites/verifier.md): Run verification checks against the current session state
- [Data Packs](https://docs.gym.scale.com/deep-dives/data-packs.md): Pre-configured datasets for website environments
- [Deep Dives](https://docs.gym.scale.com/deep-dives/overview.md): Comprehensive guides for Scale Gymnasium
- [Desktop Task Design](https://docs.gym.scale.com/deep-dives/task-design/desktop.md): Creating tasks for desktop VM environments
- [MCP Task Design](https://docs.gym.scale.com/deep-dives/task-design/mcp.md): Creating tasks for MCP tool-use environments
- [Task Design Overview](https://docs.gym.scale.com/deep-dives/task-design/overview.md): Creating effective tasks for agent evaluation
- [Website Task Design](https://docs.gym.scale.com/deep-dives/task-design/website.md): Creating tasks for web application environments
- [Troubleshooting](https://docs.gym.scale.com/deep-dives/troubleshooting.md): Common issues and solutions for Scale Gymnasium environments
- [Desktop Verifiers](https://docs.gym.scale.com/deep-dives/verifiers/desktop.md): File comparison and rules-based evaluation for desktop environments
- [MCP Verifiers](https://docs.gym.scale.com/deep-dives/verifiers/mcp.md): LLM Judge and rubric claims for tool-use tasks
- [Verifier Overview](https://docs.gym.scale.com/deep-dives/verifiers/overview.md): How verification works across Scale Gymnasium environments
- [Website Verifiers](https://docs.gym.scale.com/deep-dives/verifiers/website.md): State checks, log checks, and rubric evaluation for website environments
- [Desktop Environments Overview](https://docs.gym.scale.com/environments/desktop/overview.md): Isolated, controllable Virtual Machine desktops for CUA testing
- [MCP Environment](https://docs.gym.scale.com/environments/mcp/overview.md): 45+ Model Context Protocol servers with 300+ tools for agent interactions
- [Environment Comparison](https://docs.gym.scale.com/environments/overview.md): Compare Website, Desktop, and MCP environments to choose the right one for your use case
- [Calendr](https://docs.gym.scale.com/environments/website/calendr.md): Calendar application for scheduling and event management
- [Cloudfile](https://docs.gym.scale.com/environments/website/cloudfile.md): Cloud storage application for file storage and organization
- [Pandora's Inbox](https://docs.gym.scale.com/environments/website/email.md): Email application environment for testing email-related agent tasks
- [Website Environments Overview](https://docs.gym.scale.com/environments/website/overview.md): Full-featured synthetic websites for GUI-based agent testing
- [Shopora](https://docs.gym.scale.com/environments/website/shopora.md): Sephora-style e-commerce platform
- [Choose Your Path](https://docs.gym.scale.com/getting-started/choose-your-path.md): Decide how to use Scale Gymnasium based on your needs
- [Key Concepts Overview](https://docs.gym.scale.com/getting-started/key-concepts.md): Understanding the core terminology used in Scale Gymnasium
- [Docker Quick Start](https://docs.gym.scale.com/getting-started/quickstart-docker.md): Run Scale Gymnasium environments locally with Docker
- [Web UI Guide](https://docs.gym.scale.com/getting-started/web-ui-guide.md): Complete guide to using the Scale Gymnasium Web interface
- [Scale Gymnasium](https://docs.gym.scale.com/index.md): Standardized environments for training and evaluating LLM agents

## OpenAPI Specs

- [openapi](https://docs.gym.scale.com/api-reference/openapi.json)

## Optional

- [Scale AI](https://scale.com)